Read row x to row y in a textfile

Hi,

I have a very long textfile that I want to read the lines from 32781 to 32894, excluding the lines above and below, and store the lines in a new text files. I have made this code, but it does not work.

The txt-file have some lines with two columns and other with three and four. The lines I am interested in only have two columns.

[T1,PSA1]=textread('FFC_M7_1.txt', '%f %f','headerlines',32781);
[T,PSA] = [T1(32782:32894), PSA1(32782:32894)]
f1=fopen('FFC_M7_1_test.txt','w');
for i= 1:length(T)
fprintf(f1,'%11.4f %11.4f\n',[T(i) PSA(i)]);
end
fclose(f1);

This message appear when I try to run the code:

Error using dataread
Trouble reading floating point number from file (row 114, field 1) ==> Frequency (Hz)              A
Error in textread (line 174)
[varargout{1:nlhs}]=dataread('file',varargin{:}); %#ok<REMFF1>
Error in WritePSA (line 1)
[T1,PSA1]=textread('FFC_M7_1.txt', '%f %f','headerlines',32781);

Does anyone know how to do this?

Thanks, Marthe

Answer 1

[Cedric]

Is the following working?

 [T1,PSA1]=textread('FFC_M7_1.txt', '%f %f %*s %*s','headerlines',32781);

or

 content    = fileread( 'FFC_M7_1_test.txt' ) ;
 lineStarts = [0, strfind( content, sprintf('\n') )] + 1 ;
 block      = content(lineStarts(32782) : (lineStarts(32895)-1)) ;
 data       = reshape( str2double( regexp( block, '[\d\.\-]+', 'match' )), 2, [] ).' ;

I cannot test right now though.

[Marthe Fenne Vestly]

The first is not working, I get the same message as before.

The second give "Unexpected matlab expression" in

block      = content(lineStarts(32782) : (lineStarts(:32895)-1)) ;

[Cedric]

See my edited answer (without the :). It seems that your data block doesn't have only two columns but also text occasionally, and that's probably why the first approach doesn't work.

The second approach is less sensitive to that as it extracts only the numbers. Yet if you have numbers in extra text it will fail (the reshape). In such case, you should extract the block that you are interested in, and attach it to your question (as a text file) so we can have a look at how to deal with the content properly.

[Marthe Fenne Vestly]

Statement is incomplete for

data       = reshape( str2double( regexp( block, '[\d\.\-]+', 'match' )), 2, []

I have attached the part of the file (included one headline) that I want to extract.

[Cedric]

I am back on a MATLAB. I tested the code above with an updated version of your file (attached) where I added random content after and before a block that I am interested in, going from line 73 to line 186:

 content    = fileread( 'PSA2.txt' ) ;
 lineStarts = [0, strfind( content, sprintf('\n') )] + 1 ;
 block      = content(lineStarts(73) : (lineStarts(186)-1)) ;
 data       = reshape( str2double( regexp( block, '[\d\.\-]+', 'match' )), 2, [] ).' ;

and it works. Array data contains the data that you sent me. If the full data file is not too confidential, feel free to email it to me at [e-mail], and I can have a look.

[Marthe Fenne Vestly]

Unfortunately it did not work on the full txt-file. I have email you the full file.

[Cedric]

Here an approach based on pattern matching to first find the relevant block(s) and then parse it/them. There are in fact six blocks in the full file. If you only need the first, you can do this:

 content = fileread( 'FFC_M7_1.txt' ) ;
 data = regexp( content, 'Period .sec.  PSA .g.\s*(.*?)\s*F', 'tokens', 'once' ) ;
 data = reshape( sscanf( data{1}, '%f' ), 2, [] ).' ;

and you get data as a numeric array with the values. If you need all of them:

 content = fileread( 'FFC_M7_1.txt' ) ;
 data = regexp( content, 'Period .sec.  PSA .g.\s*(.*?)\s*F', 'tokens' ) ;
 for k = 1 : numel( data )
    data{k} = reshape( sscanf( data{k}{1}, '%f' ), 2, [] ).' ;
 end

Here data is a cell array that you can index with data{1}, data{2}, etc. Each cell (each data{k} contains a numeric array with the values.

[Marthe Fenne Vestly]

This worked, thank you so much.

[Cedric]

My pleasure! Please [Accept] the answer if I solved your problem.

[Marthe Fenne Vestly]

One more thing. I want to read the lines below (295656 to 295661) and only column 1 (depth) and 2 (PGA). Is there a way to modify the program above to do this?

    Depth (m)     PGA (g)       Min. Displacement (m)       Max. Displacement (m)
 0             0.010817     -0.0775801344                0.0564587136 
 29.99994      0.00762791   -0.0540203136                0.0464908392 
 64.9998192    0.00844498   -0.034562796   0.0401955 
 122.9999064   0.00618368   -0.01831372512               0.02322947856 
 151.9997976   0.00420221   -0.01291852128               0.01819089072 
 269.9997648   0.0146314    -0.00774551664               0.00782610576

[Cedric]

Yes, you can do it this way:

 tokens = regexp( content, 'Depth \(m\)\s+PGA.*?[\r\n]+([^D]+)', 'tokens', 'once' ) ;
 profileData = reshape( sscanf( tokens{1}, '%f' ), 4, [] ).'

With that you get:

 profileData =
         0    0.0077   -0.0340    0.0366
   29.9999    0.0059   -0.0245    0.0274
   64.9998    0.0066   -0.0191    0.0171
  122.9999    0.0049   -0.0114    0.0139
  151.9998    0.0037   -0.0093    0.0109
  269.9998    0.0101   -0.0041    0.0057

[Marthe Fenne Vestly]

I think the code extract the wrong PGA. The depth (column 1) are correct, but I can´t figure out where the PGA values comes from.

[Cedric]

We are probably not working with the same file. This is what I have:

[Marthe Fenne Vestly]

You are right, I was comparing two different files. Sorry about that, and thanks a lot.

[Cedric]

My pleasure!